Update app.py

app.py

@@ -2,11 +2,13 @@ import torch
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
+# ----------------------------
+# Model configuration
+# ----------------------------
 MODEL_NAME = "Qwen/Qwen2.5-3B-Instruct"
 
 tokenizer = AutoTokenizer.from_pretrained(
-    MODEL_NAME,
-    trust_remote_code=True
+    MODEL_NAME, trust_remote_code=True
 )
 
 model = AutoModelForCausalLM.from_pretrained(
@@ -16,6 +18,9 @@ model = AutoModelForCausalLM.from_pretrained(
     trust_remote_code=True
 )
 
+# ----------------------------
+# System prompt for CBE marking
+# ----------------------------
 SYSTEM_MESSAGE = {
     "role": "system",
     "content": (
@@ -26,21 +31,23 @@ SYSTEM_MESSAGE = {
     )
 }
 
+# ----------------------------
+# Chat function
+# ----------------------------
 def chat(user_input, messages):
     if not messages:
         messages = [SYSTEM_MESSAGE]
 
-    # Ensure clean message format
+    # Ensure all contents are strings
     messages = [
         {"role": m["role"], "content": str(m["content"])}
         for m in messages
     ]
 
-    messages.append({
-        "role": "user",
-        "content": str(user_input)
-    })
+    # Add user input
+    messages.append({"role": "user", "content": str(user_input)})
 
+    # Create Qwen prompt
     prompt = tokenizer.apply_chat_template(
         messages,
         tokenize=False,
@@ -49,9 +56,10 @@ def chat(user_input, messages):
 
     inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
 
+    # Generate response
     outputs = model.generate(
         **inputs,
-        max_new_tokens=384,   # ↓ safer for 3B
+        max_new_tokens=384,   # safe for 3B
         temperature=0.7,
         top_p=0.9,
         do_sample=True
@@ -62,22 +70,20 @@ def chat(user_input, messages):
         skip_special_tokens=True
     )
 
-    messages.append({
-        "role": "assistant",
-        "content": response
-    })
+    # Append assistant response
+    messages.append({"role": "assistant", "content": response})
 
     return messages, ""
 
+# ----------------------------
+# Gradio UI
+# ----------------------------
 with gr.Blocks() as demo:
     gr.Markdown("## 🤖 Qwen 2.5 – 3B Chatbot (English & Kiswahili)")
 
-    chatbot = gr.Chatbot()
-    msg = gr.Textbox(
-        label="Your message / Ujumbe wako",
-        autofocus=True
-    )
-    state = gr.State([])
+    chatbot = gr.Chatbot()  # old Gradio safe
+    msg = gr.Textbox(label="Your message / Ujumbe wako", autofocus=True)
+    state = gr.State([])    # keeps messages
 
     msg.submit(
         chat,
@@ -85,8 +91,9 @@ with gr.Blocks() as demo:
         outputs=[chatbot, state]
     )
 
-demo.queue(concurrency_count=1)
-
+# ----------------------------
+# Launch safely on HF Spaces
+# ----------------------------
 if __name__ == "__main__":
     demo.launch(
         server_name="0.0.0.0",